phys169@canterbury.ac.nz (09/24/90)
What affects the speed of reading data off a hard disk? That is, if I run some program that works out the optimum interleave for a computer/controller/disk combination, what could make it non-optimum (e.g. changing CPU speed, running some TSR's, etc)?? Also, it is possible to determine the best interleave without actually writing to the disk (purely from timing of reading what is there)? Mark Aitchison, Physics, University of Canterbury, New Zealand.
RFM@psuvm.psu.edu (09/24/90)
The new Norton Utilities, version 5.0, has a utility called Configure which tests verious interleave factors and optimizes your hd with the most efficient. Don't know exactly how it works ( I haven't read the manuals), but the program changed the default interleave on my Seagate ST238 (RLL) from 3 to a value of 5. Disk I/O seems faster, & Norton specs indicate it is. Wonder how caching compares with interleave for speedy I/O? Bob M., PSU-Harrisburg
ted@helios.ucsc.edu (Ted Cantrall) (09/26/90)
In article <1990Sep24.113019.9222@canterbury.ac.nz> phys169@canterbury.ac.nz writes: > >Also, it is possible to determine the best interleave without actually writing >to the disk (purely from timing of reading what is there)? SpinRite from Gibson Research can tell you what the current interleave is and what the data transfer rate would be for interleaves from 1:1 up to 8:1. It them lets you pick the one you like and -without destroying data- will reset it to that. Slick trick. -ted- ------------------------------------------------------------------------------- ted@helios.ucsc.edu |"He has showed you, O man, what is good; and what does the W (408)459-2110 |Lord require of you but to do justice and to love kindness H (408)423-2444 |and to walk humbly with your God?" Micah 6:8 (RSV)
aceverj@accucx.cc.ruu.nl (Jaap Verhage) (09/27/90)
In article <7165@darkstar.ucsc.edu> ted@helios.ucsc.edu (Ted Cantrall) writes: >In article <1990Sep24.113019.9222@canterbury.ac.nz> phys169@canterbury.ac.nz writes: >> >>Also, it is possible to determine the best interleave without actually writing >>to the disk (purely from timing of reading what is there)? >SpinRite from Gibson Research can tell you what the current interleave is >and what the data transfer rate would be for interleaves from 1:1 up to 8:1. >It them lets you pick the one you like and -without destroying data- will >reset it to that. Slick trick. -ted- You could also try hdtst128.arc from wuarchive.wustl.edu, /mirrors/msdos/dskutl. It'll do the same for you and is free. -- Regards, Jaap. Jaap Verhage, Academic Computer Centre, State University at Utrecht, Holland. aceverj@cc.ruu.nl +<-*|*->+ I claim *every*thing and speak for myself
srm@dimacs.rutgers.edu (Scott R. Myers) (09/28/90)
Okay I have something to add to this discussion. It's concerning
interleave on IDE drives. Here goes the story...
I read on the net how QEMM drivers can slow a system down which can
have an effect on hard disk throughput. The fix: Change the
interleave. Okay so what I did is ftp'd some interleave stuff from
Simtel "iau" & "spintest". Well when I ran these programs they had
problems determining my existing interleave. I kept getting messages
concerning timing errors. Can anyone explain why this is happening.
If the info I gave is vauge let me know and I'll try to post more
detail. Thanx in advance.
srm
--
Scott R. Myers
Snail: 26 Stiles Street Phone:(201)882-3100
Apartment 18
Elizabeth, NJ 07201
Arpa: srm@dimacs.rutgers.edu Uucp: ..!dimacs!srm
"... No matter where you go, there you are ..."bruce@ssc-vax.UUCP (Bruce Stock) (10/03/90)
This is not an answer to the original query, but my own experience with setting interleave recently. I ran two public domain interleave analyzer packages, both recommended a 2:1 interleave. Both required that I disable SMARTDRIVE (The Windows 3 disk cache) during the test. However, I found by actual reformating and benchmarking that an interleave of 3:1 gave me better performance when using SMARTDRIVE, which is my normal configuration. So the moral of the story is, be sure to test the config- uration you are going to run, otherwise you may not get the optimum.